Previous Frequently Asked Questions (FAQs)
faq1.pdf [PDF 20KB]
faq2.pdf [PDF 94KB]
faq3.pdf [PDF 178KB]
Patients, diagnoses and trial entry
What is Chronic Fatigue Syndrome?
Chronic Fatigue Syndrome (CFS) is an illness in which a person is disabled by severe fatigue and other symptoms, which have lasted at least six months, and there is no other disease found that could explain the symptoms.
How is CFS defined?
Several published definitions or criteria can be applied to decide if a person has CFS. We used the Oxford definition (see below) to define CFS in the PACE trial.
We also assessed participants to see if they met the International (Centers for Disease Control) definition, to see whether the effects of treatments were different in those who met this alternative definition. 67 per cent of PACE trial patients met this definition.
What is ME?
ME or myalgic encephalomyelitis/encephalopathy refers to an illness in which severe fatigue is a symptom, which is characteristically exacerbated by minimal activity, along with other symptoms that fluctuate over time.
We assessed whether patients in the PACE trial met the London criteria for ME. We did this to see whether the effects of treatment differed in those who met criteria for ME. 51 per cent of PACE trial patients met the London criteria for ME.
What is the difference between CFS and ME?
The relationship between the two conditions is controversial. However, it is generally agreed that ME is a similar or related condition to CFS; some regard it as the same condition; others believe it to be a distinct illness.
What are the “Oxford” criteria for CFS?These require that a person has had at least six months of severe fatigue, that fatigue is their main symptom, and that the fatigue is disabling and accompanied by other symptoms. It is also required that the person has no other medical condition that could explain the symptoms. All patients in the PACE trial met these criteria.
Why did you choose the Oxford criteria for defining CFS?
There is no “gold standard” definition of chronic fatigue syndrome, and around 20 definitions have been published. We chose the Oxford criteria because: (a) we wanted to find out which treatments were best in those who had fatigue as their principal symptom. Some patients, clinically diagnosed as having CFS, may have another symptom, such as pain, as their primary symptom. (b) The Oxford definition of CFS is the most straightforward to use in clinical practice and has been used in many previous research studies. We also assessed participants to see if they met the International (Centers for Disease Control) criteria for CFS.
Did you include people who did not have CFS?
The study used stringent procedures to ensure that those people with another diagnosis that would explain their fatigue were excluded. These included a full history and physical examination by a specialist doctor, ten blood tests and a urine test. All potential participants also underwent a standardised psychiatric interview. Together these procedures ensured that patients entering the trial were suffering from CFS.
Did you include people with ME?
Half (51 per cent) of trial participants met criteria for ME, as defined by the London criteria, which were based on the original description by the late Dr Melvin Ramsay, physician at the Royal Free Hospital. The trial results for people who met criteria for ME were the same as for those who did not.
The trial and treatments
What treatments did you test?
Participants were randomly allocated to one of four treatment groups. All participants received specialist medical care (SMC). One group received this alone. The other three groups also received either adaptive pacing therapy (APT), cognitive behaviour therapy (CBT) or graded exercise therapy (GET). All treatments were delivered on an individual basis.
Specialist Medical Care (SMC) consisted of seeing a doctor in a clinic that specialised in the management of CFS. This included providing general advice about managing the illness and prescribing medicines for symptoms such as sleep problems and pain, as well as encouraging self-help when SMC was provided by itself without an additional therapy.
Adaptive Pacing Therapy (APT), delivered by occupational therapists, assumed that the illness could not be changed by changing behaviour, and aimed to help the patient to use their energy wisely and to allow natural recovery, by both stabilising and balancing activities with rest, while staying within the limits imposed by the illness. All participants also received Specialist Medical Care (SMC).
Cognitive behaviour therapy (CBT) was provided by a clinical psychologist or nurse therapist. After stabilising the level of activity, CBT aimed to help the patient to do more and feel better by testing out the best ways to cope with the illness. This included a gradual return to activities and challenging what both patients and therapists had identified as potentially unhelpful ways of coping and thinking about the illness. All participants also received SMC.
Graded exercise therapy (GET) was delivered by physiotherapists and consisted of an individually tailored exercise programme. This began with stabilisation of levels of activity and then incremental increases in physical activity, agreed between participant and physiotherapist. The planned increments took into account symptoms, fitness, and current activity levels. The treatment aimed to help patients do more and feel better by gradually increasing the time they were physically active and then increasing the intensity of activity. All participants also received SMC.
Why did you select these treatments for study?
CBT and GET had been shown in previous small trials to be moderately effective treatments for patients with CFS. However, some patient organisations had expressed concern about their safety and efficacy, and had reported that patients preferred pacing or specialist medical care. We therefore wanted to find out which treatments were both safe and effective.
Is it a problem that participants were not blinded to the treatments being received?
Both the therapists and patients knew what treatments they were being given. This is unavoidable for trials of rehabilitative treatments in which the patient is an active participant.
The statistician doing the main analysis did not know which treatment group participants were in (treatment groups were labelled A, B, C or D), to avoid any potential bias in the conduct of the analysis.
Was the treatment effect due to the participants choosing a particular treatment?
The allocation of treatment to participants was random and patients could not select which treatment they received.
Was the treatment effect due to some participants getting more attention than others?
This is very unlikely as the three additional therapies were closely matched for amount of patient contact.
Was the treatment effect due to participants expecting to get better with CBT or GET?This is also very unlikely as we measured participants’ expectations of their allocated treatment, after they knew which one they were getting and before they started it. This indicated that patients expected CBT to be least likely to be helpful. Despite this, CBT emerged as one of the two treatments that helped the most. In contrast, patients about to receive adaptive pacing therapy had the highest expectations of benefit, but this was the least effective. This all suggests that the participant’s expectations of treatment did not determine the trial outcomes.
During the trial, a newsletter was distributed that contained quotes from participants. Could this have caused bias in the study?
It is considered good practice to publish newsletters for participants in trials, so that they are kept fully informed both about the trial’s progress and about news of developments in their illness.
The investigators published four such newsletters during the trial, which can all be found here, and all newsletters were approved by the independent research ethics committee before publication.
For example, in this Newsletter [PDF 395KB], there are six comments from patients about their treatments, but no specific treatment or therapy is named, and investigators were careful to print feedback from participants from all four treatment arms. We included these comments to try to encourage patients to consider entering the trial, in order to improve recruitment. However, by the time it was published, following ethical approval, all participants had been recruited. It would have been very unlikely that the newsletter could have biased participants as any influence on their ratings would affect all treatment arms equally.
The same newsletter also mentioned the publication of the UK National Institute for Health and Care Excellence guideline for the management of CFS (this institute is independent of the UK government). The guidelines were described in summary form, and the only reference made to individual therapies was in the following sentence: “The guidelines emphasise the importance of joint decision making and informed choice and recommended therapies include Cognitive Behavioural Therapy, Graded Exercise Therapy and Activity Management.” These three therapies were the ones being tested in the trial. It is therefore unlikely that it would lead to bias in the direction of one or other of these therapies
The findings <
What did the PACE trial show?
Both cognitive behaviour therapy (CBT) and graded exercise therapy (GET), when combined with specialist medical care (SMC), were more effective in reducing fatigue and improving physical functioning than either adaptive pacing therapy (APT) when combined with SMC, or SMC alone. Approximately 12 out of 20 patients made a clinically useful reduction in fatigue and improvement in functioning with either CBT or GET compared to about 8 out of 20 with APT and 9 out of 20 with SMC.
Twelve months after starting in the trial, 3 out of 10 participants were within normal population ranges for both fatigue and function, following CBT and GET, which were approximately twice as many participants than after APT and SMC. This means patients were more able to do things we all take for granted such as carrying shopping, or walking up a flight of stairs. This level of improvement is what we would expect in the treatment of other chronic disabling conditions. Being within the normal population range for these two outcomes does not necessarily mean the patient had recovered from CFS.
Have any other studies found similar results to the PACE trial?The main results of the study have been verified by a reanalysis of the trial data by a Cochrane Review group. Furthermore a number of published systematic reviews and meta-analyses have supported the findings. These include Whiting et al, 2001, Edmonds et al, 2004, Chambers et al, 2006, Malouff et al, 2008, Price et al, 2008, Castell et al, 2011, Larun et al, 2015, Marques et al, 2015 and Smith et al, 2015.
How effective were CBT and GET?
We concluded that both CBT and GET were more effective than APT and SMC and that the size of the effect was moderate. We drew these conclusions based on the size of their effect compared with APT and SMC and the proportions of participants who made a clinically useful reduction in fatigue and improvement in functioning.
Why did CBT and GET have similar effects? Is it just the effect of seeing a therapist?
Both CBT and GET were better than APT, so this suggests it was not simply due to the benefit of seeing a therapist. The better outcome may have been due to the active rehabilitative approach common to both CBT and GET, which encourages people gradually to do more. However, we have now analysed in more detail what makes these treatments work, and have reported these findings (also explained in a previous FAQ document [PDF 178KB]).
Were the treatments safe?
We measured the safety of all the treatments in five separate ways, and found no significant differences in any of these measures among the different arms of the trial. We concluded that all these treatments are safe so long as they are delivered by similar healthcare professionals, who are trained and supervised to deliver these treatments in a similar way to the PACE trial.
Are the results applicable to those worst affected?
We do not know, as we did not study housebound participants. Results cannot therefore be extrapolated to those who are severely affected.
Does adaptive pacing therapy (APT) not work?
A minority of people who received APT did improve, but no more than the proportion who received SMC by itself. The majority of patients who received APT reported that they were satisfied with their therapy, but more than any other treatment APT was more likely to lead to worse physical functioning.
How is it possible that APT had the most satisfied group of patients but was the least effective therapy?
Satisfaction with a treatment is based on lots of things, such as how helpful the person found their therapist, and does not necessarily relate to its effectiveness. Patients can be satisfied that their therapist did their best, without the therapy itself improving symptoms and disability.
Did poor delivery of APT account for the findings?
We know the therapists delivered APT to a high standard and that APT was delivered as planned because of:
1. The high rates of patient satisfaction,
2. Recorded therapy sessions were consistent with the therapy manual when independently assessed,
3. The quality of the therapeutic relationships between patients and therapists, which were independently assessed, were as good with APT as with other therapies.
This suggests that it was the content of the therapy that was less effective, not the therapists or how it was delivered. It seems as though the adaptive nature of APT is not as effective as the rehabilitative approach common to both CBT and GET.
Would there be additional non-specific effects of seeing a therapist that could have influenced the results?
The non-specific effects of a therapist can be important. However, they would only account for the different outcomes between treatments if there was a difference in how the therapies were delivered between the treatment groups. It is unlikely that this was the case in the PACE trial. We ensured that all therapists were well trained and supervised, and that the total number of treatment sessions and time offered was the same across all therapies. Patients reported that they were similarly satisfied after all three therapies (APT, CBT and GET). Independent scrutineers of audio recordings of therapies reported that the therapies had been delivered as intended. In summary, it seems improbable that the differences between the therapy groups could be due to non-specific effects.
If the treatments received after the end of the trial and before the 2.5 year follow up were no longer allocated randomly, would this make the results harder to interpret?
Yes it did. After the end of the main study (one year post-treatment) participants were able to choose further treatments if they wished, thus breaking the original randomisation. In clinical trials, it is considered unethical to deny additional treatments to patients following the main phase of a trial if they are requested or needed.
In the findings of the 2.5 year follow up study it was noted that the originally randomly allocated treatment had often been supplemented by an additional treatment, and that this would limit what we could learn by comparing patients in their originally randomised treatment groups.
However, the main findings of the long-term follow were clear – that there was no deterioration from the one year improvements in patients originally allocated to CBT and GET.
Measuring the outcomes of the trial
Why did you choose a fatigue questionnaire and a scale measuring physical function as the primary outcome measures?
After discussion with patients and a national patient charity, we decided that these measures best captured improvement in the patients’ main symptoms and ability to do things, whilst not being too long or complicated to score.
What is the Chalder fatigue questionnaire?
This 11 item scale measures different aspects of fatigue, and was scored by the trial participants to provide an overall measure of fatigue and associated symptoms; the higher the score the more severe the fatigue.
What is the SF-36 physical function scale?
This 10 item sub scale (from a scale with 36 items) measures different aspects of physical ability. It was used to provide a measure of overall physical function; the higher the score the better the physical function.
Why did you use subjective outcomes?
Following consultation with a patient charity, we chose (subjective) self-ratings as the primary outcomes, since we considered that the patients themselves were the best people to determine their own state of health.
We have also reported the results of a number of potentially more objective outcomes (although all are also dependent on subjective factors of motivation and confidence), including a walking test, a stepping test, employment status and financial benefits, in papers published in the Lancet, PLoS One and Lancet Psychiatry.
The distance participants could walk in six minutes was significantly improved following GET, compared to other treatments.
There were no significant differences in fitness, employment or welfare benefits between any of the treatments. In fact, the numbers of patients receiving welfare benefits went up in all treatment groups, perhaps because they were given a definite diagnosis, and received advice regarding eligibility for welfare benefits during the trial. Treatment manuals can be found here which include the advice regarding welfare benefits and work.
We interpreted these data in the light of their context and validity. For instance, we did not use employment status as a measure of recovery or improvement, because patients may not have been in employment before falling ill, or they may have lost their job because of being ill. Getting better and getting a job are not the same things, and being in employment depends on the prevailing state of the local economy as much as being fit for work.
Note: The above question has been answered elsewhere in 2008, 2011, 2013 (here and here) and 2015.
Could changes in the scoring of the two primary outcomes from the original trial protocol have caused bias?
The two primary outcomes for the trial in the original trial protocol were the SF-36 physical function sub-scale and the Chalder fatigue questionnaire. These were not changed.
However, as a trial like PACE takes many years to complete, discussion about how best to analyse the findings continued after it began, but before any data was analysed. These discussions led to changes in how we analysed the two primary outcomes (fatigue and physical function). The two independent oversight committees approved the changes. The detailed analysis plan, including these changes, was published, and these changes and the reasons for them were also described in the main paper.
The changes were
A. Simplifying the main analysis. Originally the analysis was going to use a more complex measure, which combined two ways to measure improvement (the number of patients who either exceeded a threshold score or who improved by more than 50 per cent). After careful consideration, it was decided that this composite measure would be very hard to interpret clinically, and would not allow the direct comparison of effectiveness between treatments. We therefore chose to compare mean (average) scores of each outcome measure. This is a better way to see which treatment works best on which outcome, and made the main findings easier to understand and interpret.
B. To use the Likert scoring method rather than the binary methods originally proposed for the fatigue questionnaire. This change was to achieve greater variability in the score and greater sensitivity to change. This is because the Likert method allows the questionnaire answers to have a score of 0, 1, 2 or 3 to indicate how much of a problem each fatigue factor is for the participant; 11 questions therefore gives a score that can range from 0 to 33. The binary method of scoring only considers a yes or no; 11 questions can therefore only produce a range of 0 to 11. Both Likert and binary methods of scoring have been described before and both are regularly used by other researchers.
There were no changes to the scoring method of the physical function scale
These changes have been explained in several publications (here and here), including explicit descriptions and justification within the main paper itself, the statistical analysis plan, and the trial website section of frequently asked questions, published in 2011 [PDF 94KB].
Were some of the patients well when they entered the trial?
All patients entering the trial met the Oxford criteria for CFS and also had scores above thresholds of severity for fatigue and physical function. Some 13 per cent had scores within the normal population range (defined as within one standard deviation of the population means) for either one or other of fatigue and physical function. This is not the same as being well or as being ‘recovered’ (see below).
After an editorial suggested that having scores within the normal population range was consistent with recovery, the authors published a correction to clarify that this was not the case.
Did the trial look at recovery?
The main trial report did not address recovery. Recovery is a complex concept that is hard to define.
We did, however, look at recovery in an exploratory secondary analysis of different criteria and explored different ways of defining it.
Why did you change the criteria for determining recovery from the original protocol in the secondary analysis?
Over the long period from writing the original protocol to planning the analysis, we reconsidered our original definition of recovery. As a result, we tried to improve how we measured the concept of recovery and the different ways it can be defined.
For instance, we included those who felt “much” (and “very much”) better in their overall health as one of the five criteria that all had to be fulfilled to define recovery.
We also changed the thresholds for judging recovery on the two primary outcomes (fatigue and physical function) from the mean (average) for the population to the population mean plus a standard deviation from the mean, to better reflect the levels of fatigue and physical function found in the general population. This was informed by a recently published population study.
These changes were all made before the analysis occurred (i.e. they were pre-specified) and were fully described and explained in the paper itself. Furthermore, in the relevant paper we discussed the difficulties in measuring recovery, and showed how other ways of defining recovery could produce different results. The bottom line was that, however we defined recovery, significantly more patients had recovered after receiving CBT and GET than after other treatments.
Why did you not use more objective measures of recovery, such as employment or the distance walked in six minutes, in the secondary analysis?
We addressed this point in 2013 in correspondence that followed the paper. In summary, we did not think that being employed was the same as being recovered, since it was likely that some participants were not working before they became ill, or that they may have lost their job because of being ill. Many factors affect whether a person gains new employment.
We also explained that we did not use the distance walked in six minutes, as we were unable to use this measure in a way comparable to other studies, due to lack of space in some of our clinics and our concern not to exacerbate ill health in patients undertaking this test. Furthermore, some 18 per cent of participants did not undertake this test at follow up.
Do the results suggest that CFS is ‘all in the mind’?
The study did not find that CFS is ‘all in the mind’ – in fact, it did not look at causes of the disease at all.
Some people may think that if treatments like CBT help CFS patients, it is saying something about the nature of the illness, but this is not true. In the main results paper it was stated: "The effectiveness of behavioural treatments does not imply that the condition is psychological in nature."
The study tells us nothing about the cause of the illness, simply how effective different treatments are for helping people who have it.
The team that designed and ran the trial include many doctors and healthcare professionals who have worked and continue to work in chronic fatigue syndrome clinics, sometimes for many years.
They are aware of the serious and debilitating nature of chronic fatigue syndrome, and have witnessed in their clinics how the illness can ruin people’s lives and affect their families. There is no doubt in their minds that the illness can last for years and is a real and chronic condition, and they do not think CFS is ‘all in the mind’ or imaginary.
Some of those involved in the PACE trial also undertake research into the biological nature of the illness; research that has indicated that there are some biological abnormalities that have been found repeatedly in CFS.
How have you found a treatment to be useful when we do not know the cause of CFS?
There are very many examples in medicine where a treatment is developed for an illness before the cause of the illness is known. For example: quinine treating malaria, or digitalis (digoxin from the foxglove) helping heart failure. Treatments sometimes help in reversing the factors that keep someone unwell rather than addressing original causes. Digitalis is an example of that. CFS and ME are extremely debilitating and therefore there is a real need for treatment now. We are open minded about what research into the condition’s causes will reveal; in the meantime the PACE trial shows that CBT and GET can make a difference to patients’ lives today.
Why did not all participants improve?
Since approximately 6 out of 10 patients improved after either cognitive behavioural or graded exercise therapies, the results of this study are encouraging. But, as with treatments in most areas of medicine, there will be patients who do not respond to treatment. This underscores the need for continued research into this area.
Does the PACE trial say that CBT and GET are the only treatments for CFS?
The trial found CBT and GET to be moderately helpful and also that they do not benefit everyone. All patients also received specialist medical care, which might involve being prescribed medicines to help symptoms such as insomnia or pain.
It has always been recognised that more research is needed to find other treatments that help. This was recently stated in the results of the 2.5 year follow up when it was noted that in all of the treatment groups some patients remained unwell at long-term follow-up. This was described as “an observation that reminds us that better treatments are still needed for patients with this chronically disabling disorder.”
Who funded the PACE trial?
The trial was funded by the Medical Research Council (MRC), the Department of Health, the Department of Work and Pensions and the Scottish Chief Scientist Office.
Would there have been any conflicts of interest, for example, the investigators’ past involvement in insurance companies?
No insurance company was involved in any aspect of the trial. There were some 19 investigators, three of whom had done consultancy work at various times for insurance companies. This consultancy work was not related to the conduct of the research. Furthermore, it is not clear how this would be a conflict of interest; other than, the interest shared by patients and health services in finding a treatment that helps patients.
Following the trend to report all interests, whether judged to be conflicting or not, all sources of income were listed as potential conflicts of interest in the relevant papers.
The patient information sheet informed all potential participants as to which organisations had funded the research, which is consistent with ethical guidelines.